Amage All in One

Amage All in One is the ComfyUI-Amage pack's "everything in one panel" node. It bundles a prompt, image size, denoise, iteration count, and three spare typed slots into a single node - then passes each value out on its own output so you can fan them across your graph. It's a settings-hub, not a generator: it doesn't render anything itself. Think of it as a poor man's version of what rgthree's Context node does for whole groups, reduced to a pile of values.

Why you'd reach for it

If you keep tweaking the same five knobs - prompt, width, height, denoise, iterations - and they're scattered across three different nodes in your workflow, this drags them all into one visible place. It makes a clean "settings region" at the top of your graph that you can edit without hunting. That's a real workflow nicety for a small personal graph. The Text category it lives in undersells it; it's really a parameter hub.

How it works - and the honest part

The mechanism is almost embarrassingly simple: it concatenates input_text + text into prompt_text (exactly like its sibling Amage Text), then every other input is passed straight through to the matching output. No math, no conversion, zero cleverness. And that's where the trap is: aspect_ratio looks like it should do something - the node even names it as if it sizes the image - but it's just a string that passes through untouched. Set it to "16:9" and nothing resizes. width and height are whatever integers you typed; nothing derives one from the other. If you expected this node to be an aspect-ratio calculator, that's not it.

The inputs that matter

Ten inputs, but you'll actually touch these:

• text - your main prompt, multiline.
• input_text - a single-line prefix, handy for a fixed style tag or LoRA trigger that stays put while you swap text.
• width / height - the numbers you'll wire into an EmptyLatentImage.
• denoise - default 0.75, feeds a KSampler's denoise for img2img passes.
• n_iter - iteration count, 1–100. Note ComfyUI's stock KSampler has no "iterations" concept, so this is only useful if you consume it yourself (a script loop or a sampler pack that honors it).

The rest - aspect_ratio, string_1, int_1, float_1 - are pass-through luggage for whatever you want to carry around the graph.

Outputs

Nine named outputs: prompt_text, width, height, aspect_ratio, denoise, n_iter, string_1, int_1, float_1. The practical wiring: prompt_text → CLIPTextEncode, width/height → EmptyLatentImage, denoise → KSampler. Keep those three and you've got a tidy little control panel.

Install

Same pack as the other Amage nodes - ComfyUI Manager, search "Amage", install, restart. Or:

cd ComfyUI/custom_nodes
git clone https://github.com/Sampkao/ComfyUI-Amage.git
# restart ComfyUI

Pack-wide gotcha: no requirements.txt, and the module imports transformers at load (used only by the Amage STT node). If transformers isn't in your ComfyUI venv, every node in the pack - including this one - fails to load. pip install transformers into the venv and you're clear.

Is it worth it? If you already run rgthree, this is redundant - its context/power nodes do more. But if you want a dead-simple, one-node knob panel and you're already installing the pack, it's a fine little convenience. Just remember the aspect ratio doesn't do anything.